REVIEW OpenAccess Timingandcarriersynchronizationin ...users.cecs.anu.edu.au/.../_papers/journals/2016_eurasip.pdfsurvey paper are those interested in or already work-ing in synchronization

Nasir et al. EURASIP Journal onWireless Communications andNetworking (2016) 2016:180 DOI 10.1186/s13638-016-0670-9

REVIEW Open Access

Timing and carrier synchronization inwireless communication systems: a survey andclassification of research in the last 5 yearsAli A. Nasir1, Salman Durrani2*, Hani Mehrpouyan3, Steven D. Blostein4 and Rodney A. Kennedy2

Abstract

Timing and carrier synchronization is a fundamental requirement for any wireless communication system to workproperly. Timing synchronization is the process by which a receiver node determines the correct instants of time atwhich to sample the incoming signal. Carrier synchronization is the process by which a receiver adapts the frequencyand phase of its local carrier oscillator with those of the received signal. In this paper, we survey the literature over thelast 5 years (2010–2014) and present a comprehensive literature review and classification of the recent researchprogress in achieving timing and carrier synchronization in single-input single-output (SISO), multiple-inputmultiple-output (MIMO), cooperative relaying, and multiuser/multicell interference networks. Considering bothsingle-carrier and multi-carrier communication systems, we survey and categorize the timing and carriersynchronization techniques proposed for the different communication systems focusing on the system modelassumptions for synchronization, the synchronization challenges, and the state-of-the-art synchronization solutionsand their limitations. Finally, we envision some future research directions.

Keywords: Timing synchronization, Carrier synchronization, Channel estimation, MIMO, OFDM

1 IntroductionMotivation: The Wireless World Research Forum(WWRF) prediction of seven trillion wireless devicesserving seven billion people by 2020 [1] sums up thetremendous challenge facing existing wireless cellularnetworks: intense consumer demand for faster data rates.Major theoretical advances, such as the use of multipleantennas at the transmitter and receiver (multiple-inputmultiple-output (MIMO)) [2, 3], orthogonal frequency-division multiple access (OFDMA) [4], and cooperativerelaying [5–7] have helped meet some of this demandand have been quickly incorporated into communicationstandards. These technologies also form a core part of thenext-generation cellular standards, 5G, which is underdevelopment [8, 9].

*Correspondence: [email protected] work was supported in part by the Australian Research Council’sDiscovery Project funding scheme (project number DP140101133).2Research School of Engineering, Australian National University (ANU), 2601Canberra, AustraliaFull list of author information is available at the end of the article

In order to fulfill the demand for higher data rates,a critical requirement is the development of accurateand realizable synchronization techniques to enable novelcommunication paradigms. Such synchronization tech-niques allow communication systems to deliver higherdata rates, e.g., through the use of higher-order mod-ulations and utilization of cooperative communicationschemes. Hence, there has been considerable researchrecently in synchronization techniques for novel commu-nication strategies.Aim: The aim of this paper is to provide a survey and

classification of the research in the field of synchroniza-tion for wireless communication systems that spans thelast 5 years (2010–2014). This is not an easy task giventhe large number of papers dealing with synchronizationand its associated challenges in both current and emerg-ing wireless communication systems. The critical needfor such a survey is highlighted by the fact that the lastcomprehensive survey paper on synchronization was pub-lished nearly a decade ago [10]. While survey papers onsynchronization for wireless standardization have recentlyappeared [11–13], these surveys do not overview the

© 2016 Nasir et al. Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 InternationalLicense (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in anymedium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commonslicense, and indicate if changes were made.

http://crossmark.crossref.org/dialog/?doi=10.1186/s13638-016-0670-9-x&domain=pdfmailto: [email protected]://creativecommons.org/licenses/by/4.0/

Nasir et al. EURASIP Journal onWireless Communications and Networking (2016) 2016:180 Page 2 of 38

state-of-the-art published research. For an overview of thestate-of-the-art in synchronization research prior to 2010,see the 2009 special issue on synchronization in wire-less communications of the EURASIP Journal on WirelessCommunications and Networking [14].In this survey, we overview the relationships within the

published research in terms of systemmodel and assump-tions, synchronization challenges, proposed methods,and their limitations. We also highlight future researchdirections and important open problems in the fieldof synchronization. The main intended audience of thissurvey paper are those interested in or already work-ing in synchronization. This survey paper aims to enableresearchers to quickly immerse themselves in the currentstate of the art in this field. Moreover, by highlightingthe important open research issues and challenges, webelieve the paper would stimulate further research inthis field. Since this paper is not intended to be a tuto-rial on synchronization, we deliberately avoid presentingmathematical details and instead focus on the big picture.Background and scope: Synchronization is a common

phenomenon in nature, e.g., the synchronized flashing offireflies or the synchronous firing of neurons in the humanbrain [15, 16]. In wireless communications, timing andcarrier synchronization are fundamental requirements[17]. In general, a wireless receiver does not have priorknowledge of the physical wireless channel or propagationdelay associated with the transmitted signal. Moreover, tokeep the cost of the devices low, communication receiversuse low-cost oscillators which inherently have some drift.In this context, timing synchronization is the process bywhich a receiver node determines the correct instants oftime at which to sample the incoming signal and carriersynchronization is the process by which a receiver adaptsthe frequency and phase of its local carrier oscillator withthose of the received signal. Particularly, depending on thespecific communication systems, synchronization defini-tion/procedure could be very different. According to 3rdGeneration Partnership Project, a terminology referredto as cell search has been widely used to represent anentire synchronization procedure, which may constituteboth initial and target cell searches. Timing synchroniza-tion may consist of frame/slot/symbol/chip synchroniza-tions, residual timing tracking, first arrival path search(in terms of OFDMA), multi-path search (in terms ofCDMA), etc. Similarly, carrier synchronization may implyinteger/fractional frequency offset estimation (in termsof OFDMA), coarse/fine frequency offset estimation (interms of CDMA), residual frequency offset tracking, etc.Major advances in timing and carrier synchroniza-

tion such as pilot-symbol-assisted modulation [18] areused in present-day cellular networks to achieve car-rier accuracy of 50 parts per billion and timing accu-racy of 1μs (±500 ns) [11]. The requirement in future

wireless networks is toward tighter accuracies, e.g., tim-ing accuracy of 200 ns, to enable location-based services[12]. Hence, there is a need for new and more accu-rate timing and carrier estimators. In general, in orderto quantify the performance of any proposed estimator, alower bound on the mean-square estimation error can bederived. The bounds are also helpful in designing efficienttraining sequences. In addition, for multiple parametersneeded, say, for the joint estimation of timing and carrierfrequency offsets, these bounds include coupling infor-mation between the estimation of these parameters. Forexample, if the bound suggests very-low coupling betweenthe estimation of timing and carrier frequency offsets,this implies that these parameters can be estimated sepa-rately without any significant loss in the estimation perfor-mance. In particular, there usually exists strong couplingbetween channel and carrier frequency offset estimationand their joint estimation is helpful to achieve improvedestimation accuracy [19, 20].Although timing and carrier synchronization are neces-

sary for successful communication, they cannot provide acommon notion of time across distributed nodes. Clocksynchronization is the process of achieving and main-taining coordination among independent local clocks toprovide a common notion of time across the network.Some wireless networks, such as worldwide interoper-ability for microwave access (WiMAX), are synchronizedto the global positioning system (GPS) [12]. Others, e.g.,Bluetooth, wireless fidelity (Wi-Fi), and Zigbee rely on abeacon strategy, where all nodes in the network follow thesame time reference given by a master node broadcast-ing a reference signal [12]. For recent surveys on clocksynchronization, please see [21–25].In the literature, timing and carrier synchronization

techniques are sometimes considered in conjunction withradio frequency (RF) front-end impairments. RF impair-ments arise as a result of the intrinsic imperfectionsin many different hardware components that comprisethe RF transceiver front ends, e.g., amplifiers, convert-ers, mixers, filters, and oscillators. The three main typesof RF impairments are I/Q imbalance, oscillator phasenoise, and high-power amplifier (HPA) nonlinearities[26]. I/Q imbalance refers to the amplitude and phasemismatch between the in-phase (I) and quadrature (Q)signal branches, i.e., the mismatch between the real andimaginary parts of the complex signal. Oscillator phasenoise refers to the noise in an oscillator, mainly due to theactive devices in the oscillator circuitry, which introducesphase-modulated noise, directly affecting the frequencystability of the oscillator [27]. The HPA nonlinearitiesrefer to the operation of the HPA in its nonlinear regionwhen working at medium- and high-power signal levels.The influence of these RF impairments is usually mit-igated by suitable compensation algorithms, which can


be implemented by analog and digital signal processing.For a detailed discussion of RF impairments, the readeris referred to [28]. In cases where RF impairments (typ-ically I/Q imbalance or phase noise) are considered inconjunction with timing and carrier synchronization, theyare identified separately in the classification.Methodology: Synchronization is generally considered

as a subfield of signal processing. According to GoogleScholar, nine out of the top ten publication avenues in sig-nal processing are IEEE journals [29]. Hence, we used theIEEEXplore database to search for papers on timing andcarrier synchronization.We selected papers (in December2014) by searching for words “frequency offset” OR “fre-quency offsets” OR “timing offset” OR “timing offsets”in IEEEXplore metadata only. In order to focus on theimportant recent advances, we limited our search to alljournal papers published in the last 5 years only, i.e., from2010 to 2014. Also, we limited the search to the followingconferences: ICC, GLOBECOM, VTC, WCNC, SPAWC,and PIMRC, because it was found that these conferencescontained sufficient numbers of papers to address thesynchronization topics.Using these principles, papers that dealt with timing and

carrier synchronization were carefully selected for inclu-sion in this survey paper. A classification of these papers,

with respect to the adopted communication system, ispresented in Table 1. Some papers were found to studythe effect of timing and carrier synchronization on theperformance of various communication systems, but theydid not directly estimate or compensate for these syn-chronization impairments. These papers are summarizedin Table 2 for the sake of completeness. However, thesepapers are not discussed in the survey sections below.Abbreviations and acronyms: The list of abbreviations

and acronyms used in this paper are detailed in Table 3.In the paper, in Tables 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14,15, 16, 17, 18, 19, 20, and 21, “CSI Req.” column indicates(using yes/no value) whether or not channel state infor-mation (CSI) is required for synchronization procedure;“CE” column indicates (using yes/no value) whether or notalgorithm considers channel estimation (CE); “Est/Comp”column indicates whether algorithm only considers esti-mation (Est) of parameters or also uses the estimatedparameters for compensating (Comp) their effect on sys-tem bit error rate (BER) performance; “N/A” stands fornot applicable; and “Bound” column indicates whether thepaper derives or provides lower bound on the estimationperformance.Organization: The survey is organized as follows. The

selected papers are classified into five categories: (i) SISO

Table 1 Classification of papers on timing or carrier synchronization, 2010–2014

Communication system Single-carrier Multi-carrier

SISO (Tables 4, 5, 6, and 7) [30–68, 323, 380–383] [66, 69–158, 384–422]

Multiple antenna (Tables 8 and 9) [160–170] [172–203]

Cooperative (Tables 10 and 11) QF-OWRN [205]

AF-OWRN [19, 20, 206, 215–218] [225, 242–247]

DF-OWRN [19, 20, 204, 206–213] [225–241]

AF-TWRN [219–224] [248–250]

DF-TWRN [214]

Multiuser/multicell

(Tables 12, 13, 14, 15, 16, 17, and 18) SC-FDMA uplink [251, 253–262]

OFDMA uplink [253, 263–292, 380, 423–430]

CDMA [294] [295–301]

Cognitive radio [306, 307, 312] [304, 305, 308–311, 313–316]

Distributed multiuser [317–321]

CoMP [326, 332] [324, 325, 327–331, 333]

Multicell interference [334, 335] [336–341]

Others (Tables 19, 20, and 21) UWB [345–349, 356] [350–355, 357, 358]

Spread spectrum [359–361] [362, 363]

mmwave [368, 369] [370, 371]


Table 2 Papers on the effect of timing or carrier synchronizationon the system performance, 2010–2014

Communication Single-carrier Multi-carriersystem

SISO [431–442] [441, 443–457]

Multiple [458, 459] [460–462]antenna

Cooperative AF-OWRN [463, 464]

TWRN [465]

Multiuser/ SC-FDMA uplink [466]multicell

OFDMA uplink [467–472]

CDMA [473]

Cognitive radio [474] [475–479]

Distributed [480]multiuser

Multicell [471, 481, 482]interference

Others UWB [483]

Spread spectrum [484] [485]

(Section 2), (ii) MIMO (Section 3), (iii) cooperative relay-ing (Section 4), (iv) multicell/multiuser (Section 5), and(v) others (ultra-wide band (UWB) and spread spectrum)communication networks (Section 6). Each category issplit into single-carrier and multi-carrier (e.g., OFDM) sys-tems. For each category, we discuss the system modelfor synchronization, the synchronization challenges, andthe state-of-the-art synchronization solutions and theirlimitations. Future research directions and importantopen problems are highlighted in Section 7. Finally,Section 8 concludes this survey.

2 SISO systems2.1 Single-carrier SISO communication systems2.1.1 SystemmodelIn single-carrier SISO systems, a single antenna transmit-ter communicates with a single antenna receiver and theinformation is modulated over a single carrier.The transmitter is assumed to communicate with

the receiver through an additive white Gaussian noise(AWGN) or frequency-flat/frequency-selective fadingchannel. At the receiver end, the effect of channel canbe equalized either in the time domain or the frequencydomain. Time domain equalization is a simple single tapor multi-tap filter. In frequency domain equalization, alsoreferred to as single-carrier frequency domain equaliza-tion (SC-FDE), frequency domain equalization is carried

out via the fast Fourier transform (FFT) and inverse fastFourier transform (IFFT) operations at the receiver.

2.1.2 Synchronization challengeThe received signal at the receiver is affected by a singletiming offset (TO) and a single-carrier frequency offset(CFO). The receiver has to estimate these parameters andcompensate for their effects from the received signal inorder to decode it. The receiver may or may not have theknowledge of CSI. In case of no CSI availability, a receiverhas to carry out CE in addition to TO or CFO estimation.The estimation of TO and CFO can be achieved usingpilots or by blind methods. For pilot-based estimation, atransmitter sends a known training sequence (TS) to thereceiver before sending the actual data. For blind estima-tion, a receiver estimates the synchronization parametersusing unknown received data. Note that there exists cou-pling between channel and CFO estimation and theirjoint estimation is helpful to achieve the best estimationaccuracy for these parameters [19, 20].

2.1.3 Literature reviewThe summary of the research carried out to achieve tim-ing and carrier synchronization in single-carrier SISOcommunication systems is given in Table 4:

1. The estimation or compensation of timing offsetalone and frequency offset alone is studied in [30–50]and [51–56], respectively.

2. Joint timing and carrier synchronization is studied in[57–67].

The categorized papers differ in terms of channelmodel, channel estimation requirements, or pilot/trainingrequirements. They also differ in whether proposing esti-mation, compensation, joint channel estimation, or esti-mating lower bound. Further details or differences amongthese papers are provided in the last column of Table 4,which further indicates whether any additional param-eter such as phase noise (PHN), IQ imbalance, signal-to-noise ratio (SNR) estimation, or direction of arrival(DoA) estimation, is considered.Moreover, whether train-ing sequence (TS) design or hardware implementation istaken into consideration is also labeled in this table.

2.1.4 SummaryTiming and carrier synchronization for single-carrierSISO communication systems is a very well-researchedtopic. The majority of the papers in Table 4 are publishedbefore 2012. Generally, it is not possible to identify thebest pilot-based and best blind-based estimators since thepapers have widely different system model assumptions.One possible future work in this area should compare theperformance of their proposed solutions to existing work


Table 3 List of common acronyms and abbreviations

Acronym Definition

AF Amplify and forward

AFD-DFE Adaptive frequency domain decision feedback equalizer

AOD Angle of departure

BER Bit error rate

CE Channel estimation

Comp Compensation

CFO Carrier frequency offset

CP Cyclic prefix

CSI Channel state information

DoA Direction of arrival

DF Decode and forward

DL Direct link

DLC-SFC Distributed linear convolutional space frequency code

DLC-STC Distributed linear convolutional space time coding

DSFBC Distributed space frequency block coding

DSTBC Distributed space time block coding

DSTC Distributed space time coding

Est Estimation

FBMC Filter bank multi-carrier

FDE Frequency domain equalization

FDMA Frequency division multiple access

FD-S3 Frequency domain-spread spectrum system

FFT Fast Fourier transform

Freq. flat Frequency flat

Freq. sel. Frequency selective

GD-S3 Gabor division-spread spectrum system

HetNet Heterogeneous network

IFFT Inverse fast Fourier transform

IFO Integer frequency offset

IQ In-phase quadrature-phase

IR Impulse radio

MAI Multiple access interference

MB-OFDM Multiband-OFDM

MC Multi-carrier

MCFOs Multiple carrier frequency offsets

MISO Multiple-input single-output

MTOs Multiple timing offsets

N/A Not applicable

OSTBC Orthogonal space time block coding

OWRN One-way relaying network

PHN Phase noise

PUs Primary users

req. Required

Rx Receiver

SC Single carrier

Table 3 List of common acronyms and abbreviations (Continued)

SCO Sampling clock offset

SDR Software defined radio

SFBC Space frequency block coding

SFCC Space frequency convolution coding

SFO Sampling frequency offset

SIMO Single-input multiple-output

STC Space time coding

TD-LTE Time division Long-Term Evolution

TH Time hopping

TR-STBC Time reversal space time block code

TO Timing offset

TS Training sequence

TWR Two-way ranging

TWRN Two-way relaying network

Tx Transmitter

UFMC Universal filtered multi-carrier

WSN Wireless sensor network

in Table 4 with similar assumptions to make clear how thestate of the art is advancing.Since it is desired to implement both estimation and

compensation of timing and carrier frequency offsets,a few research works have considered such problemsassuming AWGN [63, 64], frequency flat [61], and fre-quency selective channels [57]. The futurework in this areamay consider [57, 61, 63, 64] as baseline research workand further develop for more efficient estimation andcompensation techniques and hardware implementation.

2.2 Multi-carrier SISO communication systems2.2.1 SystemmodelIn multi-carrier systems, information is modulated overmultiple carriers. The well-known multi-carrier system isbased upon OFDM.1 In OFDM systems, at the transmit-ter side, an IFFT is applied to create an OFDM symboland a cyclic prefix is appended to the start of an OFDMsymbol. At the receiver, the cyclic prefix is removed andan FFT is applied to the receivedOFDM symbol. Note thatfrequency domain processing greatly simplifies receiverprocessing. The length of the cyclic prefix is designed tobe larger than the span of the multipath channel. Theportion of the cyclic prefix which is corrupted due to themultipath channel from the previous OFDM symbols isknown as the inter-symbol interference (ISI) region. Theremaining part of the cyclic prefix which is not affectedby the multipath channel is known as the ISI-free region.Note that cyclic prefix can mainly remove the ISI and theproper design of the cyclic prefix length has been a designissue under research.


Table 4 Summary of synchronization research in single-carrier SISO communication systems

Article Channel model CSI Req. CE Blind/pilot Est/Comp TO/CFO Bound Comments

[57] Freq. sel. No No Blind Both Both Yes FDE

[58] AWGN N/A N/A Pilot Est Both Yes TS design

[59] Freq. flat Yes No Pilot Est Both Yes

[30] AWGN N/A N/A Pilot Both CFO No Turbo coding

[31] AWGN N/A N/A Blind Comp CFO N/A Hardware implementation

[32] AWGN N/A N/A Blind Comp CFO N/A

[33] AWGN N/A N/A Blind Est CFO Yes

[34] AWGN N/A N/A Pilot Est CFO Yes


[60] Freq. sel. No No Pilot Est Both No

[36] AWGN N/A N/A Blind Both CFO Yes LDPC coding

[37] Freq. sel. Yes No Blind Both CFO N/A FDE

[38] Freq. sel. No No Pilot Both CFO No IQ imbalance

[51] AWGN N/A N/A Blind Est TO No CFO presence

[61] Freq. flat Yes No Pilot Both Both No PHN

[52] AWGN N/A N/A Blind Both TO Yes LDPC coding

[39] Freq. sel. No Yes Pilot Both CFO No PHN

[62] AWGN N/A N/A Pilot Est Both Yes TS design

[40] Freq. sel. Yes Yes Pilot Both CFO No FDE, IQ imbalance

[53] AWGN N/A N/A Blind Both TO No Turbo coding

[41] AWGN N/A N/A Pilot Est CFO No

[63] AWGN N/A N/A Pilot Both Both No

[64] AWGN N/A N/A Blind Both Both Yes

[42] AWGN N/A N/A Blind Both CFO No

[65] AWGN N/A N/A Blind Comp Both No Phase offset

[66] Freq. sel. No No Blind Comp Both No

[43] Freq. flat No Yes Blind Both CFO No

[44] AWGN N/A N/A Blind Est. CFO No Symbol rate est

[45] AWGN N/A N/A Blind Est CFO No Doppler-rate est

[46] AWGN N/A N/A Pilot Comp CFO No SNR est

[54] Freq. sel. Yes No Pilot Comp TO No FDE

[47] AWGN N/A N/A Pilot Both CFO No PHN

[48] Freq. flat No Yes Blind Both CFO No

[49] AWGN N/A N/A Blind Both CFO Yes

[55] AWGN N/A N/A Blind Both TO No

[50] AWGN N/A N/A Pilot Est CFO No

[67] Freq. flat Yes No Blind Comp Both No DoA estimation


2.2.2 Synchronization challengeIn OFDM systems, the presence of TO affects the systemperformance in a different ways as compared to single-carrier systems:

1. If the TO lies within the ISI-free region of the cyclicprefix, the orthogonality among the subcarriers is notdestroyed and the timing offset only introduces a

phase rotation in every subcarrier symbol. For acoherent system, this phase rotation is compensatedfor by the channel equalization scheme, which viewsit as a channel-induced phase shift.

2. If the TO is outside the limited ISI-free region, theorthogonality among the subcarriers is destroyed bythe resulting ISI and additional inter carrierinterference (ICI) is introduced.


Thus, the objective of timing synchronization in OFDMsystems, unlike in single-carrier systems, is to identify thestart of an OFDM symbol within the ISI-free region of thecyclic prefix.The presence of CFO in OFDM systems attenuates

the desired signal and introduces ICI since the modu-lated carrier is demodulated at an offset frequency at thereceiver side. In OFDM systems, CFO is usually repre-sented in terms of subcarrier spacings and can be dividedinto an integer part (integer number less than the totalnumber of subchannels) and a fractional part (within ± 12of subcarrier spacing). If the CFO is greater than the sub-carrier spacing, a receiver has to estimate and compensatefor both the integer and fractional parts of the normalizedCFO.The synchronization in OFDM systems can be per-

formed either in the time domain or the frequency domaindepending upon whether the signal processing is executedpre-FFT or post-FFT at the receiver, respectively.

2.2.3 Literature reviewThe summary of the research carried out to achieve car-rier synchronization, timing synchronization, and jointtiming and carrier synchronization in multi-carrier SISOcommunication systems is given in Tables 5, 6, and 7,respectively. Their details are given below.

1. Carrier synchronization:The papers studying carrier synchronization inmulti-carrier SISO communication systems arelisted in Table 5. It can be observed that there aregroups of papers which consider the same channelmodel and the same requirement for CSI andtraining. Also, they consider the same problem interms of estimation or compensation. In thefollowing, we describe how these papers differwithin their respective groups.

(a) Pilot-based CFO estimation andcompensation with channel estimation:The papers here can be grouped into twocategories. The first group does not providean estimation error lower bound [68–84]. Inaddition to carrier synchronization, [68]proposes to achieve seamless service invehicular communication and also considersroad side unit selection; [69] considers CFOtracking assuming constant modulus-basedsignaling; [70] considers concatenatedprecoded OFDM system, [71] proposesMMSE-based estimation; [72] proposeshard-decision-directed-based CFO tracking;[73] considers phase-rotated conjugate

Table 5 Summary of research in multi-carrier SISO communication systems considering carrier synchronization

Article Channel model CSI Req. CE Blind/pilot Est/Comp TO/CFO Bound

[68–84] Freq. sel. Yes Yes Pilot Both CFO No

[85–92] Freq. sel. Yes Yes Pilot Both CFO Yes

[386] Freq. sel. Yes Yes Pilot Est CFO No

[93–95] Freq. sel. Yes Yes Pilot Est CFO Yes

[96–101] Freq. sel. Yes No Blind Est CFO No

[419] Freq. sel. Yes Yes Blind Both CFO No

[406] Freq. sel. No No Blind Est CFO No

[380] Freq. sel. No No Pilot Est CFO No

[387] Freq. sel. Yes No Blind Both CFO No

[102, 103] Freq. sel. Yes No Blind Both CFO Yes

[104–117] Freq. sel. Yes No Pilot Est CFO No

[118–130] Freq. sel. Yes No Pilot Est CFO Yes

[393] Freq. sel. Yes No Pilot Both CFO Yes

[131–134] Freq. sel. Yes No Pilot Both CFO No

[418] AWGN Yes No Blind Both CFO No

[135–138] Freq. sel. Yes No Pilot Comp CFO No

[139–142] Freq. sel. Yes No Blind Comp CFO No

[412] Freq. sel. Yes Yes Pilot Comp CFO No

[389] Freq. sel. No Yes Pilot Est CFO No

[394] Freq. sel. Yes Yes Semiblind Both CFO Yes


Table 6 Summary of research in multi-carrier SISO communication systems considering timing synchronization

Article Channel model CSI Req. CE Blind/pilot Est/Comp TO/CFO Bound Comments

[390] Freq. sel. Yes Yes Pilot Both TO No DVB-T system

[396] Freq. sel. Yes Yes Pilot Est TO No Subspace based est

[397] Freq. sel. No No Blind Both TO No Subspace based est

[400] Freq. sel. No No Pilot Both TO Yes Autocorrelation-based estimation

[403] Freq. sel. No No Pilot Both TO Yes Fourth-order statistics

[405] Freq. sel. Yes Yes Pilot Both TO No ML estimation

[410] Freq. sel. No No Pilot Est TO Yes SNR estimation

[411] Freq. sel. No Yes Blind Both TO No Throughput computation

[414] Freq. sel. No No Pilot Both TO No

[416] Freq. sel. No No Pilot Est TO No

[421] Freq. sel. No No Pilot Est TO No

[422] Freq. sel. No No Pilot Est TO No Immune to CFO

transmission and receiver feedback; [74]considers hardware implementation with IQimbalance and power amplifier nonlinearity;[75] considers hexagonal multi-carriertransmission system and a doubly dispersivechannel; [76] considers maximum aposteriori expectation-maximization(MAP-EM)-based Turbo receiver; [77]considers an FBMC system; [78] considersaerial vehicular communication; [79]proposes noise variance estimation andconsiders EM algorithm; [80] considers SFOestimation; [81] considers hardwareimplementation, [82] proposes estimation ofthe CFO over a wide range of offset values;[83] considers IQ imbalance and phase noisedistortion; and [84] considers Dopplerspread in a mobile OFDM system.The second group of papers provides anestimation error lower bound on obtainingthe CFO [85–92]. In addition to carrier

synchronization, [85] considers IQimbalance, [87] proposes an extendedKalman filter (EKF)-based estimator in thepresence of phase noise, [88] proposes anML estimator and considers an FBMCsystem, [89] considers SFO estimation andsynchronization, [90] considers ML basedfrequency tracking, and [91] considers IQimbalance and its estimation.

(b) Pilot-based CFO estimation with channelestimation:The papers [93–95] fall under this category.In addition to carrier synchronization, [94]proposes computationally efficient, singletraining sequence-based least squaresestimation, [93] considers doubly selectivechannel estimation, and [95] proposes anEM-based ML estimator and also considersthe presence of phase noise.

(c) Blind CFO estimation with no channelestimation:

Table 7 Summary of research in multi-carrier SISO communication systems considering joint timing and carrier synchronization

Article Channel model CSI Req. CE Blind/Pilot Est/Comp TO/CFO Bound

[143–147] Freq. sel. Yes Yes Pilot Both Both No

[409] Freq. sel. Yes Yes Pilot Est Both No

[402] Freq. sel. Yes No Pilot Both Both Yes

[148, 149, 384] Freq. sel. Yes No Blind Est Both Yes

[150, 151] Freq. sel. Yes No Pilot Both Both No

[415] AWGN N/A N/A Pilot Both Both No

[152–155] Freq. sel. Yes No Pilot Est Both No

[156–158] Freq. sel. Yes No Pilot Est Both Yes

[385] Freq. sel. Yes No Blind Both Both No

[66] Freq. sel. No No Blind Comp Both No


The papers here can be grouped into twocategories. The first group does not providean estimation lower bound [96–101]. Inaddition to carrier synchronization, [96]considers a cognitive radio network and thealgorithm applies even if timing offset isunknown, [97] considers time-varyingchannels and Doppler frequency, [98, 99]consider constant modulus-based signaling,[100] considers cyclic correlation-basedestimation, the estimator proposed by [101]is based on minimum reconstruction error,and [92] proposes an EM based estimatorconsidering very high mobility.The second group of papers provides an errorlower bound on CFO estimation [102, 103].In addition to carrier synchronization, [102]proposes a Viterbi-based estimator and[103] proposes CFO estimation using singleOFDM symbol and provides closed-formexpression for the CFO estimate usingproperty of the cosine function.

(d) Pilot-based CFO estimation with no channelestimation:The papers here can be grouped into twocategories. The first group of papers doesnot provide an estimation error lower bound[104–117]. In addition to carriersynchronization, the CFO estimationalgorithm proposed by [104] is valid even iftiming offset and channel length is unknown;the algorithm proposed by [105] estimatesinteger frequency offset; the algorithmproposed by [106] estimates samplingfrequency offset in addition to CFO; [107]estimates IFO for OFDM-based digital radiomondiale plus system; [108] considers IQimbalance and direct-conversion receivers;[109] considers CFO tracking in digital videobroadcasting (DVB-T) system; [110]proposes ML-based estimation; [111, 113]propose IFO estimation with cell sectoridentity detection in Long-Term Evolutionsystems; [112] considers IQ imbalance andhardware implementation; [114] alsoconsiders SFO estimation; [115] proposesML-based estimation and considers thedesign of pilot pattern, the estimationalgorithm in [116] is robust to the presenceof the Doppler shift; and [117] considers IQimbalance and its estimation.The second group of papers provides an errorlower bound on CFO estimation [118–130].In addition to carrier synchronization, [118]

derives CRLB for the general case where anykind of subcarriers, e.g., pilot, virtual, or datasubcarriers, may exist; [119] considerseigenvalue-based estimation; [120] considerssubspace-based channel estimation withhardware implementation and SNRdetection; [121] considers IFO estimation and training sequence design; [122] considersGaussian particle filtering-based estimation;[123] proposes both IFO and FFOestimation while also considering IQimbalance and a direct conversion receiverstructure; [124] considers SFO estimationwhile proposing ML-based estimation; [125]proposes multiple signal classification or asubspace-based estimation method; [126]proposes an estimator based on thespace-alternating generalizedexpectation-maximization (SAGE)algorithm and considers IQ imbalance; [127]proposes SNR and noise power estimation;[128, 129] consider IQ imbalance; and [130]considers doubly selective fading channels.

(e) Pilot-based CFO estimation andcompensation with no channel estimation:The papers [131–134] fall under this category.In addition to carrier synchronization, [131]proposes training sequence design inDVB-T2 system, [132] considers frequencydomain pilot signaling, [133] also considersSFO estimation, and [134] considers IQimbalance and its estimation.

(f) Pilot-based CFO compensation with nochannel estimation:The papers [135–138] fall under thiscategory. The differences among them arethat in addition to carrier synchronization,[136] proposes repeated correlative codingfor mitigation of ICI, [135] proposes trainingsequence design, [137] considers cell identification in Long-Term Evolution (LTE) system,and [138] considers detection of primarysynchronization signal in LTE systems.

(g) Blind CFO compensation with no channelestimation:The papers [139–142] fall under thiscategory. The differences among them arethat in addition to carrier synchronization,[139, 140] propose EKF-based algorithm andspace time parallel cancelation schemes,respectively, to cancel out inter-carrierinterference due to CFO, [141] proposesreduction of peak-interference-to-carrierratio, and [142] considers IQ imbalance.


2. Timing synchronization:Compared to the categorized papers for carriersynchronization in Table 5, the categorized papersfor timing synchronization in Table 6 have greatersimilarity. The major differences are found in termsof channel estimation requirement, pilot/trainingrequirement, and lower bounds on the estimationperformance. Further details are provided in the lastcolumn of Table 6, which also indicates if anyadditional parameter, e.g., SNR estimation, isconsidered.

3. Joint timing and carrier synchronization:The papers studying joint timing and carriersynchronization in multi-carrier SISOcommunication systems are listed in Table 7. It canbe observed that there are groups of papers whichconsider the same channel model and the samerequirement for CSI and training. Also, they furtherconsider the same problem in terms of estimation orcompensation. In the following, we describe howthese papers differ within their respective groups.

(a) Pilot-based TO and CFO estimation andcompensation with channel estimation:The papers [143–147] fall under thiscategory. In addition to joint timing andcarrier synchronization, [143] considers IFOestimation while considering residual timingoffset, [144] considers hardwareimplementation, [145] considers FBMCsystem, [146] considers offset-QAMmodulation, and [147] considers decisiondirected-based estimation.

(b) Blind TO and CFO estimation with nochannel estimation:The papers [148, 149] fall under thiscategory. In addition to joint timing andcarrier synchronization, [148] considersdigital video broadcasting (DVB-T2)standard and [149] proposes ML estimationwith a time-domain preamble.

(c) Pilot-based TO and CFO estimation andcompensation with no channel estimation:The papers [150, 151] fall under thiscategory. In addition to joint timing andcarrier synchronization, [150] considerstime domain synchronous (TDS)-OFDMsystem which replaces cyclic prefix with apseudo noise (PN) and thus proposesPN-correlation-based synchronization, and[151] considers hardware implementation.

(d) Pilot-based TO and CFO estimation with nochannel estimation:

The first group of papers does not providean estimation error lower bound [152–155].The differences among them are that inaddition to joint timing and carriersynchronization, CFO estimation in [152]applies to a wide CFO range, i.e., ±1/2 thetotal number of subcarrier widths; [153]considers phase noise (PN)-sequence-basedpreamble; [154] considers digital videobroadcasting (DVB-T2) system; and [155]considers blind cyclic prefix length in theiralgorithm.The second group of papers provides anestimation error lower bound [156–158]. Inaddition to joint timing and carriersynchronization, [156] considers doublyselective channel, [157] considers hexagonalmulti-carrier transmission system, and [158]proposes training sequence design.

2.2.4 SummaryTiming and carrier synchronization for multi-carrierSISO communication systems is still an ongoing topic ofresearch, as evidenced by the large number of publishedpapers. In particular, there is a major emphasis on accu-rate CFO estimation for different types of systems andoften in conjunction with RF impairments such as phasenoise and IQ imbalance.Few recent research works have considered the prac-

tical problem of both estimation and compensation oftiming and carrier frequency offsets in the presenceof channel impairments [143–147]. Particularly, [147]considers known symbol padding (KSP)-OFDM system,[143, 144] also introduces a hardware co-simulationplatform, [145] considers an FMBC system, and [146] con-siders an OQAM-OFDM system. The authors in [143]only consider integer frequency offsets and the particularwork may be extended considering both integer and frac-tional frequency offsets. The works of [143–147] can beconsidered as a baseline reference for further extension inthe relevant system models.

3 Multi-antenna systems3.1 Single-carrier multi-antenna communication systems3.1.1 SystemmodelIn a multi-antenna wireless communication system, dataare transmitted across different channels that are modeledeither as quasi-static or time varying. The received sig-nal at an antenna is given by a linear combination of thedata symbols transmitted from different transmit anten-nas. In order to achieve multiplexing or capacity gain,independent data are transmitted from different transmitantennas.


A space-time MIMO decoder can be used to decodethe signal from multiple antenna streams. On the otherhand, in order to achieve diversity gain, the same sym-bol weighted by a complex scale factor may be sent overeach transmit antenna. This latter scheme is also referredto as MIMO beamforming [159]. Depending on the spa-tial distance between the transmit or receive antennas,which may differ for line-of-sight (LOS) and non-LOSpropagation, the antennas may be equipped with eithertheir own oscillators or use the same oscillator.Depending on the number of antennas at the transmit-

ter and the receiver, multi-antenna systems can be furthercategorized into MIMO systems, multiple-input single-output (MISO) systems, or single-input multiple-output(SIMO) systems. Further, if the antennas at the trans-mitter side are not co-located at a single device, such asystem is referred to as a distributed-MIMO system, i.e.,multiple distributed transmitters simultaneously commu-nicate with a single multi-antenna receiver.

3.1.2 Synchronization challengeIn multi-antenna systems, multiple signal streams arriveat the receive antenna from different transmit antennasresulting in multiple timing offsets (MTOs). In some spe-cial cases, multiple timing offsets actually reduce to a sin-gle timing offset, e.g., if multiple antennas are co-locatedat a single transmitter device, then the transmit filters canbe synchronized easily and the multiple signal streamsarriving at the receive antenna experience approximatelythe same propagation delay.If the transmit antennas are fed through independent

oscillators, the received signal at the receive antenna isaffected by multiple carrier frequency offsets (MCFOs)because of the existence of independent frequency offset

between each transmit antenna oscillator and the receiveantenna oscillator. On the other hand, if the transmitantennas are equippedwith a single oscillator, the receivedsignal at the receive antenna is affected by a single fre-quency offset. Thus, each receive antenna has to estimateand compensate for a single or multiple timing and fre-quency offsets, depending on the system model assump-tions including Doppler fading.In the case of distributed antenna systems, the receiver

has to estimate and compensate for multiple CFOs andmultiple TOs because each distributed transmit antennais equipped with its own oscillator and multiple signalstreams arriving at the receive antenna experience differ-ent propagation delays. Thus, in practice, the number ofdistributed antennas may need to be limited to avoid syn-chronization and pilot overhead associated with obtainingmultiple CFOs and TOs.

3.1.3 Literature reviewThe summary of the research carried out to achieve tim-ing and carrier synchronization in single-carrier multi-antenna communication systems is given in Table 8:

1. The estimation or compensation of CFO alone isstudied in [160–167].

2. The joint timing and carrier synchronization isstudied in [168–170].

The categorized papers differ in terms of channel model,channel estimation requirement, or pilot/training require-ment. They also differ in terms of proposing estimation,compensation, joint channel estimation, or estimationlower bound. Further details or differences are providedin the last column of Table 4, which indicates if any

Table 8 Summary of synchronization research in single-carrier multi-antenna communication systems

Article System Fading CSI Req. CE Blind/pilot Est/Comp TO/CFOBound Oscillators (Tx/Rx) Comments

[160] Virtual MIMO Freq. flat Yes No Pilot Both CFO No Single/multiple Codebook design

[161] SC-FDMAMIMO

Freq. sel. Yes No Pilot Both CFO No Single/single SFBC, PHN

[162] MIMO Freq. sel. No Yes Pilot Both CFO No Multiple/multiple FDE

[163] MIMO Freq. flat No Yes Pilot Both CFO Yes Multiple/multiple

[168] MIMO Freq. flat No No Blind Comp Both N/A Single/single STBC

[169] MISO Freq. flat No No Blind Comp Both N/A Single/single STBC

[164] SC-FDEMIMO

Freq. sel No Yes Pilot Both CFO No Single/single IQ imbalance

[165] SIMO Freq. flat No No N/A Comp CFO N/A Single/single

[166] MISO WSN Freq. sel No Yes Pilot Est CFO No Single/single AOD est.

[170] distributedMIMO

Freq. flat No Yes Blind Both Both Yes Multiple/single

[167] MISO Freq. flat No No Blind Comp CFO N/A Multiple/single STBC


additional parameter, e.g., phase noise (PHN), IQ imbal-ance, or direction of arrival (DoA) estimation, is con-sidered or if space-time block coding (STBC), spacefrequency block coding (SFBC), or codebook design isconsidered.

3.1.4 SummarySynchronization in single-carrier multi-antenna commu-nication systems has not received as much attention com-pared to synchronization in multi-carrier multi-antennacommunication systems. This may not be surprising sincethe latter is adopted in current wireless cellular standards.Still, single-carrier multi-antenna communication has gotits importance in microwave backhaul links [171]. Theimportant problem of joint estimation and compensa-tion of MTOs and MCFOs has been considered in [170].The authors assume pilot-free systems to propose blindsynchronization. Generally, performance improvement isexpected in the presence of training-based estimation andcompensation. This, along with hardware implementationof the relevant algorithms, can be the subject of possiblefuture extensions.

3.2 Multi-carrier multi-antenna communication systems3.2.1 SystemmodelIn multi-carrier multi-antenna systems, the informationat each antenna is modulated over multiple carriers. Thus,apart from the IFFT/CP addition and CP removal/FFToperations at each transmit and receive antennas, respec-tively, the system model for multi-carrier multi-antennacommunication systems is similar to the one describedfor single-carrier multi-antenna systems presented inSection 3.1.1.

3.2.2 Synchronization challengeSimilar to single-carrier multi-antenna systems, the sig-nal arriving at the receive antenna can potentially beaffected by multiple TOs and multiple CFOs, when thetransmit antennas are fed by different oscillators and aredistant from one another. Due to multiple carriers, thepresence of multiple TOs and multiple CFOs results instrong ISI and ICI. The synchronization challenge is tojointly estimate and compensate for the effect of mul-tiple TOs and multiple CFOs in order to mitigate ISIand ICI and decode the signal from multiple antennastreams.

3.2.3 Literature reviewThe summary of the research carried out to achieve tim-ing and carrier synchronization in single-carrier multi-antenna communication systems is given in Table 9:

1. The estimation or compensation of TO and CFOalone is studied in [172–174] and [175–198],respectively.

2. The joint timing and carrier synchronization isstudied in [199–203].

The number of oscillators considered by differentpapers at the transmitter and receiver, respectively, aregiven under the “Tx/Rx Oscillator” column. The catego-rized papers differ in terms of channel model, channelestimation requirement, or pilot/training requirement.They also differ in proposing estimation, compensation,joint channel estimation, or estimation lower bound.Further details or differences are provided in the lastcolumn of Table 9, which indicates if additional param-eters, e.g., phase noise or IQ imbalance, are consideredor if STBC, SFBC, coding, or hardware implementation isconsidered.

3.2.4 SummaryCompared to the estimation of single TO and singleCFO, estimation of MTOs and MCFOs is more challeng-ing, due to pilot design issues, overhead, pilot contam-ination problem, complexity, and non-convex nature ofoptimization problems. Considering multiple oscillatorsat the transceiver, joint estimation of MTOs and MCFOshas been considered in [200, 203] only. Particularly, theauthors in [203] propose a compact TS design. How-ever, both papers, [200, 203], do not consider channelestimation as it may help in improving the estimationperformance of synchronization impairments, and fur-ther, they do not suggest algorithms for compensation ofMTOs and MCFOs. Though joint estimation and com-pensation of timing and carrier frequency offset has beenstudied, e.g., in [199, 202], however, their system modelconsider single oscillator at the transceiver. The short-comings in the above key papers can be the subject ofpossible future research.

4 Cooperative communication systemsIn cooperative communication systems, the informationtransmission between the two communicating nodes isaccomplished with the help of an intermediate relay. Letus assume a general scenario with the presence of multi-ple relays. There are two important types of cooperativecommunication networks:

• One-way relaying network (OWRN), whereinformation transmission occurs in one direction viaintermediate relays.

• Two-way relaying network (TWRN), whereinformation transmission occurs simultaneously inboth directions and both nodes exchange theirinformation with the help of intermediate relays.

The relays themselves can operate in different modes.The twomost commonmodes are (i) decode-and-forward


Table 9 Summary of synchronization research in multi-carrier multi-antenna communication systems

Article System Fading CSI Req. CE Blind/Pilot Est/Comp TO/CFO Bound Tx/Rx Oscillator Comments

[199] MIMO Freq. sel. No Yes Pilot Both Both Yes Single/single

[175] MISO Freq. sel. Yes No N/A Comp CFO N/A Multiple/single Alamouti coding

[176] MISO Freq. sel. Yes No Pilot Est CFO No Single/single IFO est.

[177] MISO Freq. sel. Yes No Pilot Est CFO No Single/single IFO est.

[178] MIMO Freq. sel. No No Blind Est CFO No Single/single

[200] MIMO Freq. sel. No No Pilot Est Both No Multiple/multiple

[179] MIMO Freq. sel. No Yes Pilot Est CFO No Multiple/multiple

[180] MIMO Freq. sel. No Yes Pilot Est CFO Yes Multiple/multiple Algebraic STC

[181] MIMO Freq. sel. No Yes Pilot Both CFO Yes Single/single insufficient CP

[182] MIMO Freq. sel. No Yes Blind Both CFO Yes Single/single

[183] MIMO Freq. sel. No Yes Pilot Both CFO No Multiple/multiple Time-varyingchannel

[184] MIMO Freq. sel. No Yes Pilot Est CFO Yes Single/single

[185] MIMO Freq. sel. No Yes Blind Est CFO Yes Single/single

[186] Coded MIMO Freq. sel. No Yes Pilot Both CFO Yes Single/single Doubly sel. channel

[187] Coded MIMO Freq. sel. No Yes Semiblind Both CFO No Single/single

[172] MIMO Freq. sel. No Yes Pilot Est TO No Single/single

[188] MIMO Freq. sel. No Yes Pilot Both CFO Yes Multiple/single

[201] MIMO Freq. sel. No No Pilot Both Both No Single/single Hardwareimplementation

[202] MIMO Freq. sel. No Yes Pilot Both Both Yes Single/single

[189] MIMO Freq. sel. No Yes Pilot Est CFO Yes Single/single IQ imbalance

[173] MIMO Freq. sel. No Yes Pilot Both TO No Single/single

[190] distributed MISO Freq. flat Yes No N/A Comp CFO N/A Multiple/single Alamouti coding

[191] MIMO Freq. sel. No Yes Pilot Comp CFO No Single/single SFBC, IQ imbalance

[192] MIMO Freq. sel. No Yes Pilot Est CFO No Single/single IQ imbalance, PHN, SFO

[193] MIMO Freq. sel. Yes No Semiblind Both CFO No Single/single

[174] MIMO Freq. flat No Yes Pilot Est TO Yes Single/single

[194] MIMO Freq. sel. No Yes Pilot Comp CFO No Single/single

[195] coded MIMO Freq. sel. No Yes Pilot Est CFO Yes Single/single STBC

[203] distributedMIMO

Freq. sel. No No Pilot Est Both No Multiple/single TS design

[196] MIMO Freq. sel. No Yes Pilot Both CFO Yes Single/single TS design, IQ imbalance

[197] MIMO Freq. sel. Yes No Pilot Both CFO No Single/single

[198] MIMO Freq. sel. Yes Yes Pilot Est CFO No Multiple/multiple Time-varying channel

(DF) and (ii) amplify-and-forward (AF) operation. In DFmode, the relays decode the received signal and forwardthe decoded signal to the intended destination node(s).In AF mode, the relays do not decode the receivedmessage and simply amplify and forward the receivedsignal.In the following subsections, we review the recent liter-

ature that deals with timing and carrier synchronizationin single-carrier and multi-carrier cooperative communi-cation systems.

4.1 Single-carrier cooperative communication systemsSince the relaying operations are different in DF and AFcooperative communication systems, so too are their syn-chronizationmethodologies. In AF, for example, the relaysmay not be required to convert the received passbandsignal to baseband and to perform carrier synchroniza-tion [20]. Generally, the synchronization parameters tobe estimated differ in OWRNs and TWRNs. In TWRN,there is self-interference, which affects the way how thesynchronization problem is formulated. In the following


subsections, we provide separate literature reviews forsynchronization in AF-OWRN, AF-TWRN, DF-OWRN,and DF-TWRN for single-carrier communication sys-tems.

4.1.1 Decode-and-forward one-way relaying networkSystemmodel The communication generally takes placein two phases. During the first phase, the source trans-mits the information to the relays. During the secondphase, the relays decode the received signal and forward itto the destination. Typically, it is assumed that the directcommunication link between the source and the destina-tion is absent or blocked due to some obstacles. However,in general, there could be a direct communication linkbetween them. In such a case, the destination also hearsthe source message during the first phase and coherentlycombines it with the message received during the secondphase.

Synchronization challenge In DF-OWRN, during thefirst phase of the two-phase communication process, thesynchronization between the source and the relays orbetween the source and the destination (in the presenceof direct link) is achieved by estimating and compensat-ing for a single TO and CFO between the source andeach relay or between the source and the destination (inthe presence of direct link). During the second phase, thesynchronization between the relays and the destination isachieved by estimating and compensating for the multi-ple TOs and multiple CFOs between the multiple relaysand the destination. Note that in the case of a single relay,only a single TO and a single CFO are required to be esti-mated and compensated for at the destination during thesecond communication phase. Increasing the number ofrelays raises the challenge of pilot design and estimationoverhead.

Literature review The summary of the research car-ried out to achieve timing and carrier synchronization insingle-carrier DF-OWRN is given in Table 10:

1. Estimation or compensation of timing offsets aloneand frequency offsets alone is studied in [204–206]and [19, 207–209], respectively.

2. Joint timing and carrier synchronization is studied in[20, 210–213].

Further details or differences are provided in the lastcolumn of Table 10, which indicates whether STBC, SFBC,or training sequence (TS) design is considered.

4.1.2 Decode-and-forward two-way relaying networkSystem model TWRNs allow for more bandwidth effi-cient use of the available spectrum since they allowfor simultaneous information exchange between the two

nodes. In TWRNs, it is usually assumed that there isno direct communication link between the two nodes.During the first phase of the two-phase communicationprocess, the information arrives at the relays from thetwo nodes. The signals from the two nodes are superim-posed at the relays. During the second phase, the relaysdecode the exclusive OR (XOR) of the bits from thereceived superimposed signal and then broadcast a sig-nal constructed from the XOR of the bits back to the twonodes [214].Synchronization challenge The synchronization chal-lenge during the first communication phase of DF-TWRNis unlike that for DF-OWRN. In DF-TWRN, the relaysreceive the superimposed signals from the two nodesduring the first communication phase. Thus, unlike DF-OWRN, the received signal at each relay during the firstcommunication phase is a function of two TOs and twoCFOs, which need to be jointly estimated and compen-sated for in order to decode the modulo-2 sum of thebits from the two nodes. The synchronization challengeduring the second communication phase of DF-TWRN issimilar to that described for DF-OWRN in Section 4.1.1.Another challenge for TWRN is the pilot design in thepresence of self-interference at the relay node.

Literature review A summary of the research carried outto achieve timing and carrier synchronization in single-carrier DF-TWRN is given in Table 10. There is only onepaper in the last 5 years that falls in this category and pro-poses joint estimation and compensation of timing offsets[214]. Most of the research has considered AF relaying forTWRN due to its implementation advantages.

4.1.3 Amplify-and-forward one-way relaying networkSystem model The system model for AF-OWRN is sim-ilar to that of Section 4.1.1 for DF-OWRN. However,instead of DF operation, the relays amplify and forwardthe source information.

Synchronization challenge The synchronization chal-lenge for AF-OWRN is quite similar to that described forDF-OWRN in Section 4.1.1, except for one important dif-ference. In DF-OWRN, the signal is decoded at the relayand the received signal at the destination is impaired bymultiple TOs and multiple CFOs only between the relaysand the destination. In AF-OWRN, the received signal atthe destination is affected not only by the multiple TOsand multiple CFOs between the relays and the destina-tion (as in the case of DF-OWRN) but also by the residualTOs and CFOs between the source and the relays. Thisis due to imperfect synchronization during the first com-munication phase, and the amplified and forwarded signalfrom the relays is a function of the residual TOs and CFOsbetween the source and the respective relays.2


Table 10 Summary of synchronization research in single-carrier cooperative communication systems

Article Network DF/AF DL Fading CSI Req. CE Blind/pilot Est/Comp TO/CFO Bound Comments

[205] OWRN QF Yes Freq. flat No Yes Pilot Est TO Yes

[19] OWRN Both No Freq. flat No Yes Pilot Both CFO Yes

[217] OWRN AF No Freq. flat No Yes Pilot Both Both Yes DSTBC

[20] OWRN Both No Freq. flat No Yes Pilot Both Both Yes

[215, 216] OWRN AF No Freq. flat No Yes Pilot Both TO Yes

[218] OWRN AF Yes Freq. sel. No No N/A Comp Both No DSTBC

[206] OWRN Both No Freq. flat No Yes Pilot Est TO Yes

[209] OWRN DF No Freq. sel. No Yes Pilot Est CFO Yes

[204] OWRN DF No Freq. flat No Yes Pilot Both TO Yes TS design

[208] OWRN DF No Freq. flat Yes No N/A Comp CFO No DSFBC

[211] OWRN DF No Freq. flat Yes No N/A Comp Both No DSTC

[212] OWRN DF No Freq. flat Yes No N/A Comp Both No DLC-STC

[210] OWRN DF No Freq. flat No Yes Blind Both Both Yes

[207] OWRN DF No Freq. sel. No Yes Pilot Both CFO No

[213] OWRN DF No Freq. flat No No Pilot Comp Both No DSTBC

[221] TWRN AF No Freq. sel. No Yes Pilot Both TO No

[222] TWRN AF No Freq. flat No Yes Pilot Both TO Yes

[223] TWRN AF No Freq. flat No Yes Semi-blind Both TO No

[219] TWRN AF No Freq. flat No No Pilot Both CFO No

[220] TWRN AF No Freq. flat No Yes Pilot Both Both Yes

[224] TWRN AF No Freq. flat Yes No N/A Comp CFO N/A

[214] TWRN DF No Freq. flat No Yes Pilot Both TO No

Literature review The summary of the research car-ried out to achieve timing and carrier synchronization insingle-carrier AF-OWRN is given in Table 10:

1. Estimation or compensation of timing offset aloneand frequency offset alone is studied in [206, 215, 216]and [19], respectively.

2. Joint timing and carrier synchronization is studied in[20, 217, 218].

4.1.4 Amplify-and-forward two-way relaying networkSystem model During the first phase, similar to DF-TWRN, the relays receive the superimposed signal fromthe two nodes. During the second phase, the relaysamplify and broadcast the superimposed signal back tothe two nodes [214].

Synchronization challenge In AF-TWRN, when therelays receive the superimposed signals from the twonodes during the first communication phase, each relayonly needs to carry out timing synchronization, i.e.,estimate and compensate for the TOs between thetwo nodes and the relay [219, 220]. The reason will

be explained shortly. During the second communica-tion phase, the relays amplify and broadcast the time-synchronized version of the superimposed signal. Eachnode then needs to estimate and compensate for theMTOs between the relays and itself and the sum of themultiple CFOs from the other node-to-relays-to-itself.Note that each node in this case does not need to esti-

mate and compensate for the multiple CFOs from itselfto relays to the other node because the effect of CFOsbetween itself and the relays during the first communica-tion phase is canceled by the effect of CFOs between therelays and itself during the second communication phasedue to the use of the same oscillators [219, 220]. Due tothis very reason, the authors in [219, 220] propose to onlyperform timing synchronization at the relay nodes duringthe first communication phase in AF-TWRN.

Literature review The summary of the research car-ried out to achieve timing and carrier synchronization insingle-carrier AF-OWRN is given in Table 10:

1. The estimation or compensation of timing offsetalone and frequency offset alone is studied by[221–223] and [219, 224], respectively.


2. The joint timing and carrier synchronization isstudied by [220].

4.1.5 SummaryIn single-carrier cooperative communication systems, fewrecent research works have considered the importantproblem of both estimation and compensation of MTOsand MCFOs in the presence of channel impairments[20, 210, 217, 220]. Particularly, synchronization inOWRNs is studied by [20, 210, 217], where [210] proposesblind synchronization with blind source separation andrelay selection and [20, 217] proposes training-basedsynchronization. AA Nasir et al. [20] proposed ML-based computationally complex compensation algorithm,and the shortcoming was overcome in the follow-upwork [217] where MMSE-based efficient compensationalgorithm was developed. Finally, synchronization inTWRNs is studied by [220]. There are still many openresearch problems to solve in this area, e.g., trainingdesign in the presence of MTOs and MCFOs or hardwareimplementation of the proposed algorithms.

4.2 Multi-carrier cooperative communication systemsThe following subsections review the literature for syn-chronization in AF-OWRN, AF-TWRN, DF-OWRN, andDF-TWRN for multi-carrier communication systems.Since most of the papers consider orthogonal frequencydivision multiplexing (OFDM) as a special case of multi-carrier communication system, the system model andsynchronization challenge below are presented for OFDMsystems.

4.2.1 Decode-and-forward one-way relaying networkSystemmodel Apart from the IFFT/CP addition and CPremoval/FFT operations at the transmitter and receiverside at each node, respectively, the system model forDF-OWRN for multi-carrier systems is similar to thatdescribed for single-carrier DF-OWRN presented inSection 4.1.1.

Synchronization challenge The synchronization chal-lenge during the first communication phase between thesource and the relays is similar to that presented for SISOmulti-carrier systems in Section 2.2.2. During the sec-ond communication phase, the relays decode the receivedsignal and forward it to the destination. Thus, the result-ing signal at the destination is affected by multiple TOsand multiple CFOs resulting in strong ISI and ICI. Thesynchronization challenge is to jointly estimate and com-pensate for the effect of multiple TOs and multiple CFOsin order to mitigate ISI and ICI.

Literature review The summary of the research car-ried out to achieve timing and carrier synchronization insingle-carrier DF-OWRN is given in Table 11:

1. Estimation or compensation of frequency offsetalone is studied in [225–237].

2. Joint timing and carrier synchronization is studied in[238–241].

Further details or differences are provided in the lastcolumn of Table 11, which indicates if STBC or SFBC isconsidered.

4.2.2 Decode-and-forward two-way relaying networkSystem model Other than the IFFT/CP addition and CPremoval/FFT operations at the transmitter and receiverside at each node, respectively, the system model for DF-TWRN for multi-carrier systems is similar to that forsingle-carrier systems presented in Section 4.1.2.

Synchronization challenge The received signal at eachrelay during the first communication phase is a func-tion of two TOs and two CFOs, which need to be jointlyestimated and compensated. Thus, the synchronizationchallenge is similar to that described for the second com-munication phase of DF-OWRN with two TOs and twoCFOs. Moreover, the synchronization challenge duringthe second communication phase of DF-TWRN is alsosimilar to the one described for the second communica-tion phase of DF-OWRN in Section 4.2.1.

Literature review To the best of our knowledge, nopaper in the last 5 years falls into this category, since theresearch in the synchronization of multi-carrier TWRNhas considered AF relaying.

4.2.3 Amplify-and-forward one-way relaying networkSystem model The system model for AF-OWRN formulti-carrier systems is similar to the one for single-carrier systems presented in Section 4.1.3. However, beingan OFDM system, there are IFFT/CP addition and CPremoval/FFT operations at the source transmitter andthe destination receiver, respectively. Note that unlikeDF-OWRN for multicarrier systems, the FFT and IFFToperations are not usually conducted at the receiver andtransmitter of the relays, respectively.

Synchronization challenge The synchronization chal-lenge during the first communication phase for AF-OWRN is similar to the one described for DF-OWRNin Section 4.2.1. During the second communicationphase, similar to AF-OWRN in single-carrier systems,the received signal at the destination is affected by notonly the multiple TOs and multiple CFOs between therelays and the destination, but also by the residual TOsand CFOs between the source and the relays, which cancause additional ISI and ICI. The effects of ISI and ICI arerequired to be mitigated to achieve synchronization.


Table 11 Summary of synchronization research in multi-carrier cooperative communication systems

Article Network DF/AF DL Fading CSI Req. CE Blind/pilot Est/Comp TO/CFO Bound Comments

[247] OWRN AF No Freq. sel. No Yes Pilot Est TO No Ranging method

[246] OWRN AF No Freq. sel. No Yes Pilot Comp CFO No

[243] OWRN AF No Freq. sel. Yes No N/A Comp CFO N/A SFCC

[244] OWRN AF No Freq. sel. Yes No N/A Comp CFO N/A Alamouti coding

[245] OWRN AF Yes Freq. sel. No Yes Pilot Est CFO Yes PHN

[225] OWRN Both No Freq. sel. No Yes Pilot Both CFO No OSTBC

[242] OWRN AF Yes Freq. sel. Yes No N/A Comp CFO N/A

[226] OWRN DF Yes Freq. sel. Yes No Pilot Est CFO No

[238] OWRN DF No Freq. sel. Yes No Pilot Both Both No

[239] OWRN DF No Freq. sel. Yes No Pilot Est Both No

[227] OWRN DF No Freq. sel. Yes No Pilot Both CFO No SFBC

[228] OWRN DF No Freq. sel. Yes No N/A Comp CFO N/A SFBC

[229] OWRN DF No Freq. sel. No Yes Pilot Est CFO Yes

[230] OWRN DF No Freq. sel. No Yes Pilot Est CFO No

[231] OWRN DF No Freq. sel. Yes No Pilot Both CFO No Alamouti coding

[232] OWRN DF No Freq. sel. Yes No N/A Comp CFO N/A STC


[234] OWRN DF No Freq. sel. Yes No N/A Comp CFO N/A

[235] OWRN DF No Freq. sel. No Yes Pilot Both CFO No

[236] OWRN DF No Freq. sel. Yes No N/A Comp CFO N/A DLC-SFC




[248] TWRN AF No Freq. sel. No No Pilot Est CFO Yes TS design

[249] TWRN AF No Freq. sel. No Yes Pilot Both CFO No

[250] TWRN AF No Freq. sel. Yes No N/A Est CFO N/A

Literature review The summary of the research car-ried out to achieve timing and carrier synchronization insingle-carrier AF-OWRN is given in Table 11. The estima-tion or compensation of timing offset alone and frequencyoffset alone is studied in [225, 242–246] and [247], respec-tively. Further details or differences are provided in thelast column of Table 11, which indicates if any additionalparameter, e.g., phase noise (PHN) estimation, is consid-ered or if STBC or any other type of space coding isconsidered.

4.2.4 Amplify-and-forward two-way relaying networkSystemmodel Apart from the IFFT/CP addition and CPremoval/FFT operations at the transmitter and receiver,respectively, the system model for AF-TWRN for multi-carrier systems is similar to the one for single-carriersystems presented in Section 4.1.4. Note that unlike DF-TWRN for multicarrier systems, the FFT and IFFT oper-ations are not usually conducted at the receiver andtransmitter of the relays, respectively.

Synchronization challenge The required parameters tobe estimated and compensated to achieve synchroniza-tion in multi-carrier AF-TWRN are similar to the onedescribed in the synchronization challenge of single-carrier AF-TWRN in Section 4.1.4. The difference is tomitigate the effect of ISI and ICI in OFDM systems.

Literature review The summary of the research car-ried out to achieve timing and carrier synchronization insingle-carrier AF-OWRN is given in Table 11. The estima-tion or compensation of frequency offset alone is studiedin [248–250]. The categorized papers differ in the sensethat training sequence design is proposed in [248], jointCFO estimation and compensation with channel estima-tion is studied in [249], and CFO estimation alone isproposed in [250].

4.2.5 SummaryIn multi-carrier cooperative communication systems, theproblem of joint estimation and compensation of MTOs


and MCFOs has been considered in very few works, e.g.,[228]. Future research investigations may help to achieveefficient algorithms. In addition, solution to the particu-lar problem in TWRNs and the derivation of CRLBs areopen research areas in this field. Moreover, most of thesolutions are pilot based. Hence, it is a challenging openproblem to design semiblind and blind estimators.

5 Multiuser/multicell interference networks5.1 SC-FDMA uplink communication systemsSingle-carrier frequency division multiple access (SC-FDMA) is an extension of SC-FDE to accommodate mul-tiple users. In SC-FDMA uplink communication systems,multiple users communicate with a single receiver andthe effect of channel distortion is equalized in frequencydomain at the receiver. Like the SC-FDE receiver, FFTand IFFT modules are present in the SC-FDMA receiver.However, the SC-FDMA transmitter also incorporatesFFT and IFFT modules. Disjoint sets of M subcarriersare assigned to each of the K users, and data symbolsfrom each user are modulated over a unique set of sub-carriers through an M-point FFT operation. Next, KM-point IFFT is applied to transform the signal to timedomain, such that the output of FFT is applied to theuser-specified M inputs of IFFT block and 0 is appliedto the remaining (K − 1)M inputs of IFFT block. Next,cyclic prefix is appended at the start of the transmis-sion block to mitigate the multipath channel effect. Atthe receiver side, following KM-point FFT and frequencydomain equalization, K of M-point IFFT operations areapplied to decode the information from the K users. SC-FDMA can also be interpreted as a linearly precodedOFDMA scheme, in the sense that it has an additionalFFT processing step preceding the conventional OFDMAprocessing.Due to the presence of an independent oscillator at each

transmitting user and due to different propagation delaysbetween each user and the receiver, the received signal inthe SC-FDMA uplink communication system suffers frommultiple CFOs and multiple TOs. The combined effectof TOs and CFOs results in both ISI and loss of orthog-onality among subcarriers, which, in turn, generates ICIand multiple access interference (MAI) at the receiver.Thus, in order to decode information from each user, areceiver has to estimate multiple TOs and multiple CFOsto compensate the effect of ICI, ISI, and MAI. In uplinkcommunication systems, there is another synchronizationchallenge that the correction of one user’s frequency andtiming at the receiver can misalign those of the otherusers [251].It must be noted that LTE systems [252] use SC-

FDMA for communication in the uplink. Synchronizationis achieved through periodically transmitted primary andsecondary synchronization signals from the base station.

Any user who has not yet acquired the uplink synchro-nization can use the primary and secondary synchro-nization signals (transmitted by the base station) to firstachieve synchronization in the downlink. Next, compen-sation for the propagation loss is made as part of theuplink random access procedure [252].The research carried out to achieve timing and car-

rier synchronization in SC-FDMA uplink communicationsystems is summarized in Table 12:

1. The estimation or compensation of multiple CFOsalone is studied in [251, 253–261]. Though[253–255, 258–261] consider the same channelmodel with known CSI in order to propose multipleCFO compensation, they differ in the followingcharacteristics. Different channel allocationstrategies and their effects on interference due toCFO is studied in [253]. An interferenceself-cancelation scheme to compensate for the effectof CFO is proposed by [254].

2. Joint timing and carrier synchronization is studied in[262]. Further, [255, 260] consider multiple antennasat the users and receiver, [260] proposes blindbeamforming assuming that the multipath delay isgreater than the cyclic prefix length, and [261]proposes MMSE-FDE. Finally, joint timing andcarrier synchronization in SC-FDMA systems isproposed in [262].

Considering pilot-free transmission, joint estimationand compensation of timing and carrier frequency offsetsis studied in [262]. The work can be extended to considerpilot-based systems in order to improve the estimationaccuracy. Further, CRLB derivation in the presence of syn-chronization impairments and hardware implementationcan also be the subject of future work.

5.2 OFDMA uplink communication systemsAn OFDMA communication system is an extension ofan OFDM system to accommodate multiple users. InOFDMA uplink communication systems, multiple userscommunicate with a single receiver. Unlike OFDM sys-tems, where information of a single user is modulated overall subcarriers, in OFDMA uplink transmitter, each usertransmits over a set of assigned subcarriers. For each user,a group of M modulated symbols is applied to the user-specified M inputs of IFFT block and 0 is applied to theremaining (K − 1)M inputs of IFFT block. In the end,cyclic prefix is appended and information is transmittedfrom every user. At the receiver side, following cyclic pre-fix removal and FFT operation, equalization is carried outto decode the information from the K users.Similar to SC-FDMAuplink communication system, the

received signal in SC-FDMA uplink suffers from multiple


Table 12 Summary of synchronization research in SC-FDMA uplink communication systems

Article Fading CSI Req. CE Blind/pilot Est/Comp TO/CFO Bound Comments

[262] Freq. sel. No Yes Blind Both Both No

[253] Freq. sel. Yes No N/A Comp CFO N/A Channel Allocation, OFDMA

[254] Freq. sel. Yes No N/A Comp CFO N/A Interference Cancelation

[255] Freq. sel. No No N/A Comp CFO N/A MIMO

[256] Freq. sel. No No Blind Both CFO N/A

[251] Freq. sel. Yes No Pilot Both CFO No

[257] Freq. sel. No Not req. Blind Comp CFO No MIMO AFD-DFE

[258] Freq. sel. No No N/A Comp CFO N/A

[259] Freq. sel. No No N/A Comp CFO N/A

[260] Freq. sel. No No N/A Comp CFO N/A Multi-antenna users and blind beamforming

[261] Freq. sel. No No N/A Comp CFO N/A MMSE-FDE

CFOs and multiple TOs. The combined effects of TOsand CFOs results in ISI, ICI, and MAI. The receiver hasto estimate and compensate the effect of multiple TOsand multiple CFOs in order to decode information fromeach user. Similar to SC-FDMA uplink communicationsystems, the correction of one user’s frequency and timingat the receiver can misalign the other users.The research to achieve timing and carrier synchro-

nization in OFDMA uplink communication systems issummarized in Table 13. Almost all categorized papersstudy carrier synchronization only, [263–290], except theauthors in [291] consider ranging scheme to propose tim-ing estimation in OFDMA uplink communication system.In addition, the authors in [292] provide solutions forsynchronizing both the timing and frequency errors ofmultiple unsynchronized users in OFDMA-based spec-trum sharing system. Further extension may considerthe derivation of CRLBs in the presence of impairmentsand hardware implementation design. Overall, though thetabulated papers consider the same channel model, theydiffer with respect to providing joint channel estimation,requiring pilot/training, or providing lower bound. Fur-ther details or differences are provided in the last columnof Table 13.

5.3 Code division multiple access communicationsystems

Code division multiple access (CDMA) communicationsystems allow multiple transmitters to send informa-tion simultaneously to a single receiver. All users sharethe same frequency and time resources. To permitthis to be achieved without undue interference, CDMAemploys spread-spectrum technology, i.e., either direct-sequence spread-spectrum (DS-CDMA) or frequencyhopping spread spectrum (FH-CDMA):

• In DS-CDMA, a unique (orthogonal or nearlyorthogonal) spreading code is assigned to eachtransmitter. A modulated signal for each user isspread with a unique spreading code at thetransmitter. Finally, at the receiver, information foreach user is dispread by using the same uniquedespreading code. All users share the full availablespectrum. Note that multi-carrier CDMA(MC-CDMA) spreads each user signal in thefrequency domain, i.e., each user signal is carried overmultiple parallel subcarriers and the spreading codesdiffer per subcarrier and per user.

• FH-CDMA uses a short-term assignment of afrequency band to various signal sources. At eachsuccessive time slot of brief duration, the frequencyband assignments are reordered. Each user employs aPN code, orthogonal (or nearly orthogonal) to all theother user codes, that dictates the frequency hoppingband assignments.

Due to MUI, timing and carrier synchronization inCDMA is very challenging since compensating TO andCFO for one user may cause synchronization loss to otherusers. Particularly in CDMA systems, the presence of TOsand CFOs destroys the orthogonality of the spreadingcodes. The design of special spreading and despreadingcodes that are robust to synchronization errors is also achallenging task. Moreover, since the received signal pow-ers from different users may vary significantly, the correctsynchronized despreading may be challenging. Synchro-nization also faces multipath delay spread since PN chipshave a short time duration. InMC-CDMA systems, due tothe presence of CFOs, orthogonality among the subcarri-ers is lost and ICI is generated. Thus, the synchronizationchallenge is to estimate multiple TOs and multiple CFOs,compensate forMUI and ICI, and design proper spreading


Table 13 Summary of synchronization research in OFDMA uplink communication systems

Article Fading CSI Req. CE Blind/Pilot Est/Comp TO/CFO Bound Comments

[263] Freq. sel. Yes No Pilot Both CFO No MIMO, Rao-Blackwellized particle filter

[264] Freq. sel. Yes No Pilot Both CFO No CFO tracking

[265] Freq. sel. Yes No Pilot Est CFO No Canonical particle swarm optimization scheme

[266] Freq. sel. Yes No Pilot Est CFO Yes Alternating projection method

[267] Freq. sel. Yes No Blind Est CFO Yes Interleaved OFDMA

[268] Freq. sel. Yes No Blind Both CFO No Tile-based OFDMA

[269] Freq. sel. Yes Yes Blind Est CFO No Interleaved OFDMA

[270] Freq. sel. Yes No Pilot Both CFO Yes Conjugate gradient method

[271] Freq. sel. Yes No Pilot Est CFO Yes Interleaved OFDMA

[272] Freq. sel. Yes No Pilot Both CFO No Multiuser interference cancelation-based algorithm

[273] Freq. sel. Yes No Blind Est CFO No LS estimation

[425] Freq. sel. Yes No Pilot Est CFO Yes SFO estimation

[426] Freq. sel. Yes No Blind Est CFO Yes Precoded OFDMA

[427] Freq. sel. Yes No Pilot Both CFO No OFDM and FBMC

[253] Freq. sel. Yes No N/A Comp CFO N/A Channel allocation, SC-FDMA

[429] Freq. sel. Yes Yes Pilot Est CFO Yes Iterative ML est

[430] Freq. sel. Yes No Pilot Est CFO No Ranging method

[274] Freq. sel. Yes No Pilot Est CFO Yes Tile-based OFDMA

[275] Freq. sel. Yes Yes Pilot Both CFO Yes Doubly selective channel

[276] Freq. sel. Yes No Pilot Est CFO Yes Conjugate gradient method

[277] Freq. sel. Yes No N/A Comp CFO N/A Interleaved OFDMA

[278] Freq. sel. Yes No N/A Comp CFO N/A Multistage interference cancelation

[279] Freq. sel. Yes No N/A Comp CFO N/A Successive interference cancelation

[280] Freq. sel. Yes Yes Pilot Est CFO No Partial FFT modulation

[281] Freq. sel. Yes No N/A Comp CFO N/A Subcarrier allocation algorithm

[282] Freq. sel. Yes No Pilot Est CFO Yes MIMO OFDMA, Bayesian estimation

[283] Freq. sel. Yes Yes Pilot Both CFO No High mobility

[284] Freq. sel. Yes Yes Pilot Both CFO No Common CFO estimation

[285] Freq. sel. Yes Yes Blind Est CFO No Interleaved OFDMA, DoA estimation

[286] Freq. sel. Yes No N/A Comp CFO N/A Exploit time domain inverse matrix

[287] Freq. sel. Yes No Pilot Both CFO No Ad hoc network

[288] Freq. sel. Yes No N/A Comp CFO N/A Soft interference cancelation

[289] Freq. sel. Yes No N/A Comp CFO N/A LS and MMSE-based compensation

[290] Freq. sel. Yes Yes Pilot Both CFO Yes Low complexity estimation

[292] Freq. sel. Yes Yes Pilot Both Both No OFDMA spectrum sharing system

[291] Freq. sel. Yes Yes Pilot Both TO Yes Ranging sheme

and despreading codes to decode information from eachuser.Compared to DS-CDMA, the synchronization

challenge is slightly reduced in FH-CDMA, becauseFH-CDMA synchronization has to be within a fraction ofa hop time. Since spectral spreading does not use a veryhigh hopping frequency but rather a large hop-set, thehop time will be much longer than the DS-CDMA system

chip time. Thus, an FH-CDMA system allows for a largersynchronization error [293]. Moreover, in FH-CDMA,there is no need for synchronization among user groups,only between transmitter and receiver within a group isrequired.Table 14 summarizes the research carried out to achieve

timing and carrier synchronization in CDMA multiusercommunication systems. All categorized papers consider


Table 14 Summary of synchronization research in CDMA communication systems

Article SC/MC Fading CSI Req. CE Blind/pilot Est/Comp TO/CFO Bound Comments

[294] SC Freq. flat Yes No Blind Est CFO No

[295] SC Freq. sel. Yes No N/A Comp CFO N/A Spreading code design

[296] MC Freq. sel. Yes No Pilot Comp CFO No Multiuser detector

[297] MC Freq. sel. Yes No N/A Comp CFO N/A Multiuser detector

[298] MC Freq. sel. No Yes Blind Both CFO No STBC MIMO

[299] MC Freq. sel. Yes No Pilot Comp CFO No

[300] MC Freq.sel. Yes No Blind Both CFO No

[301] MC Freq. sel. No Yes Pilot Comp CFO N/A

DS spread spectrum technology and study the carriersynchronization alone [294–301], where as identified inthe second column of Table 14, [294, 295] considersingle-carrier communication and [296–301] considermulti-carrier communication. In addition, as detailedin Table 14, the categorized papers differ with respectto channel model, providing joint channel estimation,requiring pilot/training, proposing estimation or compen-sation of multiple CFO, or providing lower bound. Notethat further details or differences are provided in the lastcolumn of Table 14.The joint estimation and compensation of carrier fre-

quency offsets has been considered in [298, 300]. Theirwork can be extended to include timing synchronization.

5.4 Cognitive radio-based communication systemsCognitive radio networks allow unlicensed secondaryusers (SUs) access to the spectrum of the licensed pri-mary users (PUs), without impairing the performance ofthe PUs. Depending on the spectrum access strategy, thereare three main cognitive radio network paradigms [302]:

• In the underlay cognitive networks, SUs canconcurrently use the spectrum occupied by a PU byguaranteeing that the interference at the PU is belowsome acceptable threshold [303]. Thus, SUs mustknow the channel gains to the PUs and are alsoallowed to communicate with each other in order tosense how much interference is being created to thePUs.

• In the overlay cognitive networks, there is tightinteraction and active cooperation between the PUsand the SUs. Thus, SUs use sophisticated signalprocessing and coding to maintain or improve the PUtransmissions while also obtaining some additionalbandwidth for their own transmission.

• In interweave cognitive networks, the SUs are notallowed to cause any interference to the PUs. Thus,SUs must periodically sense the environment todetect spectrum occupancy and transmitopportunistically only when the PUs are silent [302].

In the context of interweave cognitive networks, SUssense the spectrum to detect the presence or absenceof PUs and use the unoccupied bands while maintaininga predefined probability of missed detection. Differentmethods are used to detect the presence of P

Documents

REVIEW OpenAccess Timingandcarriersynchronizationin ...users.cecs.anu.edu.au/.../_papers/journals/2016_eurasip.pdfsurvey paper are those interested in or already work-ing in synchronization